Journal of Experimental Marine Biology and Ecology 403 (2011) 14–20

Contents lists available at ScienceDirect

Journal of Experimental Marine Biology and Ecology

j ourna l homepage: www.e lsev ie r.com/ locate / jembe

Phylogeography of the shanny Lipophrys pholis (Pisces: Blenniidae) in the NE Atlanticrecords signs of major expansion event older than the last glaciation

S.M. Francisco a,b,⁎, C. Faria a, W. Lengkeek c, M.N. Vieira b, E.M. Velasco d, V.C. Almada a

a Eco-Ethology Research Unit, ISPA, Rua Jardim do Tabaco 34, 1149-041 Lisboa, Portugalb Departamento de Biologia Animal, Faculdade de Ciências da Universidade do Porto, Rua do Campo Alegre, 4169-007 Porto, Portugalc Marine Ecology Group, Bureau Waardenburg, Varkensmarkt 9, 4101 CK Culemborg, The Netherlandsd Instituto Español de Oceanografía, Centro Oceanográfico de Gijón Av Príncipe de Asturias 70, 33212 Gíjon, Spain

⁎ Corresponding author at: Eco-Ethology Research Un34, 1149-041 Lisboa, Portugal. Tel.: +351 218811700; f

E-mail address: sara_francisco@ispa.pt (S.M. Francis

0022-0981/$ – see front matter © 2011 Elsevier B.V. Adoi:10.1016/j.jembe.2011.03.020

a b s t r a c t

a r t i c l e i n f o

Article history:Received 25 June 2010Received in revised form 29 March 2011Accepted 31 March 2011Available online 8 May 2011

Keywords:European marine fishEvolutionary historyPatterns of genetic diversity in the NE AtlanticPopulation geneticsRock intertidal

The study of the phylogeography of inshore fish from West Europe is revealing diverse geographical anddemographical patterns. Some species conform to the phylogeographic patterns typical of terrestrialorganisms, with marked signatures of the last glaciation and a decline of genetic diversity to the north of thespecies range. Other species, however, reveal no decline in diversity with latitude and signatures ofexpansions older than the last glaciation. The shanny Lipophrys pholis is a common intertidal resident fish inwest European rocky shores. It is unable to leave the rocky stretch where it settled as a juvenile, so thatdispersal depends entirely on the planktonic larval stage. These life-history and behavioural traits make theshanny an interesting species for phylogeographical analysis, as long-range movements by adults, whichcould blur historical signals, are absent. In this paper the phylogeography of L. pholis was studied using afragment of the mitochondrial control region and one from the first intron of the S7 ribosomal protein gene.The European samples (ranging from SW Spain to the Netherlands) did not display population differentiation,isolation-by-distance or latitudinal declines in genetic diversity. Iberia was proposed as having operated asthe main glacial refugium for the shanny. The genealogy of the European population showed that the largestexpansion detected was older than the last glaciation, with lineages persisting from the early Pleistocene,which does not conform to colonisation by a few founders in the current interglacial. It is argued that if fisheshave very large population sizes and high dispersal rates, populations can efficiently track climatic shifts sothat little or no genetic structure remains after each range expansion and latitudinal gradients of geneticdiversity tend to be weak or non-existent.

it, ISPA, Rua Jardim do Tabacoax: +351 218860954.co).

ll rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

The North-eastern Atlantic and its adjacent water bodies (Medi-terranean, North and Baltic Seas) are known to have suffered dramaticand periodic changes during the Pleistocene (Briggs, 1996). Duringthe glacial peaks ice-cover and polar waters moved to much lowerlatitudes than those that prevail during interglacials such as thecurrent one that has persisted since 10 thousand years ago (kya)(Climap Project Members, 1984). Thus, much of the northern andwestern European waters must have been periodically devoid of mosttemperate organisms with possible exception of a few unglaciatedrefugial locations (e.g. Mäkinen andMerilä, 2008). In the interglacials,temperate and especially warm-temperate organisms must haverecolonised the west European shores from refugia in unglaciatedareas (e.g. Gysels et al., 2004). Locations of these putative refugia have

been proposed in warmer parts of theMediterranean (e.g. Domingueset al., 2005), to the south of the Iberian Peninsula (e.g. Francisco et al.,2009), and for cooler water species in unglaciated areas in northwestIberia, southwest Ireland and the northern Britanny — Hurd Deepregion of the English Channel, where a marine lake persisted duringthe Last Glacial Maximum (LGM — 18 kya) (e.g. Hoarau et al., 2007).

Phylogeographic histories of west European fish do not conform toa single pattern, with some species forming a single population (e.g.Petromyzonmarinus in Almada et al., 2008; Thunnus thynnus in Bremeret al., 2005) and others displaying isolation-by-distance and/or clearpopulation differentiation (e.g. Amblyraja radiate in Chevolot et al.,2006; and Atherina presbyter in Francisco et al., 2009). In some casesgenetic diversity drops steeply to the extreme north of the speciesrange (e.g. Pomatoschistus microps in Gysels et al., 2004), while inothers it displays a similar level across the species distribution (e.g.Diplodus sargus in Domingues et al., 2007a,b). Finally, the genealogiesobtained vary from very shallow (likely due to extreme bottlenecksand expansion after the last glaciation) (e.g. Almada et al., 2008) toones that can be traced back to earlier phases of the Pleistocene (e.g.Sprattus sprattus in Debes et al., 2008).

Table 1Diversity measures for the collecting sites of Lipophrys pholis for CR and S7: number ofsequences (N), number of haplotypes (Nh), percentage of private haplotypes (%Ph),number of polymorphic sites (S), haplotype diversity (h), nucleotide diversity (π) andmean number of pairwise differences (PD).

Sampling site N Nh %Ph S h π PD

CR Total 188 136 85% 103 0.993 0.032 12.03Neeltje Jans 19 19 79% 43 1.000 0.028 10.61Carantec 24 23 74% 57 0.996 0.033 12.40Gijon 20 17 56% 40 0.984 0.030 11.39Camariñas 21 18 67% 42 0.981 0.029 10.88Cabo-do-Mundo 27 24 75% 51 0.992 0.033 12.49Estoril 31 27 70% 61 0.991 0.031 11.78Lagos 31 26 58% 52 0.989 0.031 11.95Cádiz 15 15 73% 37 1.000 0.028 10.59

S7 Total 314 26 58% 25 0.663 0.005 1.58Neeltje Jans 26 5 0% 8 0.609 0.005 1.55Carantec 36 5 0% 8 0.582 0.005 1.43Gijon 42 9 15% 13 0.668 0.005 1.71Camariñas 50 11 25% 12 0.682 0.006 1.52Cabo-do-Mundo 42 11 27% 8 0.725 0.005 1.65Estoril 40 8 13% 6 0.668 0.005 1.61Lagos 44 9 22% 7 0.650 0.004 1.47Cádiz 34 8 25% 11 0.697 0.006 1.64

15S.M. Francisco et al. / Journal of Experimental Marine Biology and Ecology 403 (2011) 14–20

The shanny Lipophrys pholis (Linnaeus 1758) is a rocky intertidalfish, extremely common along the west European shores. It rangesfrom Norway to Mauritania and from the Azores and Madeira to theentrance of the Mediterranean [sporadic records which do notcorrespond to stable populations (Almada et al., 2001)]. The demersaleggs are guarded by the male and the larvae hatch ready to startexogenous feeding, living in the plankton for ca. 29 days at atemperature of 15.5–17.5 °C (Faria et al., 2002). Juveniles and adultsshow restrictedmovements, and are thought to stay for life in the samerocky stretch (Faria et al., 1996). This life-history and behavioural traitmakes the shanny an interesting species for phylogeographical analy-sis, as long-range movements by adults, which could blur historicalsignals, are absent.

A study by Stefanni et al. (2006) showed that fish from the Azoresare sister to a group that included all fish from west Europe and thesingle specimen available from Madeira. These data also suggestedthat there is no genetic structure along west Europe, but both thenumber of individuals sequenced and the number of samplinglocations were too few to allow a population genetic analysis (12individuals in Portugal and two in Scotland). The only phylogeographicanalysis (Francisco et al., 2006) failed to detect population differen-tiation along the Portuguese shore. The scope of this paper was,however, restricted to a few hundred kilometres and only mitochon-drial DNA was analysed.

In the present paper, we extend the previous study in two ways:1) we supplement the control region fragmentwith a nuclear one (thefirst intron of the S7 ribosomal protein gene) and 2) we extend thesampling coverage from near the entrance of theMediterranean to theNetherlands, thus covering almost the entire range where the shannyis known to breed. The following questions where addressed: 1) Isthere any phylogeographical structure in L. pholis along the Europeanshore?; 2) How does the level of genetic diversity vary with latitude?;and 3) How old are the signatures of expansion/contraction eventsretained in the genealogies?

2. Materials and methods

2.1. Sampling and DNA sequencing

Specimens of L. pholis were collected from eight sites along theEuropean Atlantic coast (Fig. 1, Table 1). Samples were collected in

Fig. 1. Map of sampling locations for Lipophrys pholis. Ob — Oban, Ne — Neetlje Jans, Ply — PEstoril/Avencas, Am — Amoreira, Lag — Lagos, Cad — Cadiz, Cga — Cabo-da-Gata, and Mad

rocky tide-pools and a small piece of dorsal fin was clipped andpreserved in 96% ethanol.

Total genomic DNA extraction was performed with the REDEx-tract-N-mp kit (Sigma-Aldrich). The conditions for the amplificationswere as follows: amplification in a 20 μl volume containing approx-imately 20 ng of genomic DNA extract, 0.3 μl Taq DNA polymerase(Sogent F-taq), 2 μl (10 pmol) of each primer, 3 μl of buffer (suppliedby the manufacturer) and 1 μl (2.5 mM) of each dNTP. A fragmentfrom the mitochondrial (mtDNA) control region (CR) was amplifiedusing the primers L-pro1 (5′ ACTCTCACCCCTAGCTCCCAAAG 3′) andH-DL1 (5′ CCTGAAGTAGGAACCAGATGCCAG 3′) (Ostellari et al.,1996). Polymerase Chain Reaction (PCR) was performed using thefollowing profile: an initial denaturing step of; 35 cycles of denaturingat 95 °C for 1′, annealing for 1′ at 58 °C, and extending at 72 °C for 1′30″; a final extending step of 10′ at 72 °C. In addition, the first intron ofthe S7 ribosomal protein gene (S7) was amplified with the primersS7RPEX1F (5′-TGGCCTCTTCCTTGGCCGTC-3′) and S7RPEX2R (5′-AACTCGTCTGGC TTTTCGCC-3′) (Chow and Hazama, 1998). The

lymouth, Car — Carantec, Gij — Gijon, Cam — Camariñas, CM — Cabo-do-Mundo, SP —

— Madeira.

16 S.M. Francisco et al. / Journal of Experimental Marine Biology and Ecology 403 (2011) 14–20

amplification process was as follows: one step of 3′ at 95 °C, 35 cyclesof [94 °C for 45 s, 58 °C for 45 s and 72 °C for 1 min], and a finalextending step of 10′ at 72 °C. Purification and sequencing wereperformed as services by Macrogen (Seoul, Korea) and Stabvida(Oeiras, Portugal). Details on these procedures may be requested fromthe authors. Sequences obtained were deposited in GenBank(GU196686-784, GU301928-2243 and GU983860-1) and additionalsequences were retrieved from a previous work: DQ154169-257(Francisco et al., 2006) and AY966020-33 (Stefanni et al., 2006) (TableS1). The sequences were aligned with Clustal X2 (Larkin et al., 2007).For the S7 both strands of the same specimen were recoveredfollowing the approach of Sousa-Santos et al. (2005). This approachtakes advantage of the presence of indels in a given nuclear marker,and uses them to accurately reconstruct the individual haplotypeswithout the need of probabilistic estimation.

2.2. Phylogenetic and network analyses

The software PAUP* 4.0b (Swofford, 2000) was used to constructMaximum-Parsimony (MP) using two sequences of Lipophrys trigloidesas the outgroup (AY098860 and GU301926-27, for the CR and S7respectively).

For both DNA fragments, haplotype networks were built with thesoftware TCS 1.21 (Clement et al., 2000) using the parsimony methodof Templeton et al. (1992). Due to the graphical complexity of the CRhaplotype network, an additional network was constructed with areduced dataset containing only transversions.

2.3. Demographic analyses

ARLEQUIN 3.11 (Excoffier et al., 2005) was used to estimate levels ofgenetic diversity, gene flow and population structure — by analysis ofmolecular variance AMOVA (Excoffier et al., 1992) and pairwise Φst.The correlation between geographic distance and Φst was computedwith the Mantel test (Mantel, 1967; Smouse et al., 1986) (10,000permutations; geographic distances measured along the shore line).The same software was used to test population expansion with Fu's Fs(Fu, 1997) and Tajima's D (Tajima, 1989). In addition, the fit of themismatch distribution to that expected under demographic expan-sionwas also analysed (Rogers and Harpensding, 1992; Rogers, 1995).For S7, ARLEQUIN was also used to perform the linkage disequilibrium(LD) test and exact probability tests for deviations from the Hardy–Weinberg equilibrium (HWE).

A MCMC approach taking into account phylogenetic relationshipsamong haplotypes as implemented in LAMARC 2.1.3 (Kuhner, 2006)wasused to estimate effective population size (Nef) and the exponentialgrowth parameter (g) (details on runs may be requested from theauthors). In order to compute estimates of effective population size,their changes with time, and the age of populations, we used twodivergence rates for CR: 10% (Bowen et al., 2006 between Centropygelineages) and 3.6% (Domingues et al., 2005 for the sister speciesChromis chromis and C. limbata) and one for S7: 0.46% (Bernardi andLape, 2005 for Anisotremus taeniatus and A. verginicus, two trans-isthmian sister species of blennids separated 6.0–5.3 Mya before theclosure of the isthmus of Panama). Values are divergence rates permillion years and were halved to get an estimate of mutation rates perlineage permillion years. Faria et al. (1996) reported that both sexes ofL. pholismaturewhen they are one year old in the Portuguese coast, butMilton (1983), in Ireland, reported that maturation is achieved after aperiod of two years. Thus, a maturity period of two years was assumedbearing in mind that growth is going to be relatively slow duringnortherly populations and cooler climatic periods.

Past population dynamics of the shannypopulationwere estimatedwith Extended Bayesian skyline plots (EBSP) (Heled and Drummond,2006) using BEAST v.1.6 (Drummond and Rambaut, 2007). The EBSPgenerates a posterior distribution of Nef through time using a MCMC

sampling. For both the CR and S7 the MCMC analysis was run for50×106 generations (sampled every 1000 iterations), of which thefirst 10% was discarded as burn-in (5 runs). The substitution modelsused were GTR+G for CR and HKY+G for S7, respectively. For bothmarkers, the time for most recent common ancestor (tMRCA) and themedian and corresponding credibility intervals of the EBSP weredepicted using Tracer v.1.4 (Rambaut and Drummond, 2007).

3. Results

For the CR a fragment of 380 bp was amplified and the 188sequences obtained define 136 haplotypes. A total of 103 polymorphicsites and 277 invariant positions were found. Differences amonghaplotypes corresponded to 102 transitions, 22 transversions and 2indels. For the S7 the 314 sequences (157 individuals) define a total of26 haplotypes. For this marker the fragment obtained was 646 bplong, including 25 polymorphic sites and 621 invariant positions.Differences among haplotypes corresponded to 24 mutations (9transitions, 8 transversions and 9 indels). From the 25 polymorphicsites found along the European coast twowere clearly in heterozygotedeficit, leading to a significant deviation from the HWE (pb0.001). Asfor the LD test 13.67% of the possible loci pairs are in linkagedisequilibrium, a relatively low value.

For both fragments, the genetic diversity indices showed littlevariation among collection sites and were generally high (Table 1).For the CR most haplotypes were private, with no consistent variationwith latitude. Although Neeltje Jans is the most northern location, it isthe one with the highest percentage of private haplotypes (79%). Forthis fragment there were 20 shared haplotypes representing 15% ofthe total. The haplotype that spans the greatest latitudinal range wasshared by three locations (Neeltje Jans, Lagos and Cádiz). The resultswere different for S7, as there were no private haplotypes for thesampling locations north of the Iberian Peninsula. In fact, for thisfragment, the percentage of private haplotypes was significantlylower than for CR (Wilcoxon test Pb0.05; overall: 58% vs 85%). For S7,13 haplotypes were shared among the locations (50% of total).

3.1. Phylogenetic and network analyses

For both markers the phylogenetic analyses failed to reveal anygeographical structuring along the European shore (trees not shown).

All haplotypes for the CR of the shanny were linked together on asingle Statistical Parsimony Network (SPN), under the confidencelimit of 95% (Templeton et al., 1992), with the exception of anhaplotype from Estoril (SP17) (Fig. S1). A simplified SPN built onlywith transversions is shown in Fig. 2A. The network represents a deepgenealogy with multiple levels of diversification, with some haplo-types reaching 10 mutational steps from the ancestral one (outgroupweight 0.0599). The ancestral haplotype gave rise to haplotypes foundin all geographical areas (at one mutational step). Several haplotypes,including the ancestral one, are in the centre of star-like patterns thatspan one or two mutations from the respective centre, which couldrepresent diversification processes after the last glaciation. These starsare wide spread geographically and, as stated above, represent onlythe most recent diversification events on a much deeper genealogy. Itis interesting to note that the haplotype belonging to the onlyMediterranean fish is very close to the ancestral one (Fig. 2A). For theS7 the SPN obtained is much less diverse (Fig. 2B), fact easilyexplained by the difference in mutation rate. Although simpler, thisnetwork also lacks geographical structure. It presents a star-like shapearound the ancestral haplotype (outgroup weight 0.3711) but, evenso, some of the branches reach a maximum of 10 mutational stepsfrom the ancestor, which is consistent with the CR data about thelikely persistence of several lineages in multiple glacial cycles.

Fig. 2. Statistical parsimony network of Lipophrys pholis. Abbreviations of sampling locations as in Fig.1 A) Haplotypes for theMitochondrial control region. SPN constructed only withtransversions B) Haplotypes for the first intron of the S7 ribosomal protein gene.

Table 2Geneflow among collecting sites of Lipophrys pholis represented by Nm (abovediagonal) and ΦST (bellow diagonal), and based on the Tamura-Nei model. Significantvalues of probability P are shownwith an *. Cad— Cádiz, Lag— Lagos, SP— Estoril, CM—

Cabo-do-Mudo, Cam — Camariñas, Gij — Gijon, Car — Carantec, and Ne — Neeltje Jans.Inf — Infinite.

Cad Lag SP CM Cam Gij Car Ne

CRCad 7.352 3.936 14.366 3.027 3.452 5.640 23.168Lag 0.064* inf inf 313.617 inf inf infSP 0.113* −0.010 80.849 inf inf inf 41.197CM 0.034 −0.004 0.006 105.949 73.097 inf 244.591Cam 0.142* 0.002 −0.026 0.005 inf inf 14.601Gij 0.126* −0.005 −0.015 0.007 −0.026 inf 12.680Car 0.081* −0.013 −0.025 −0.004 −0.018 −0.013 infNe 0.021 −0.003 0.012 0.002 0.033 0.038 −0.007

S7Cad 9.950 135.075 21.877 60.954 inf 11.391 infLag 0.048 inf inf inf 90.043 inf infSP 0.004 −0.012 inf inf inf inf infCM 0.022 −0.013 −0.017 inf inf inf infCam 0.008 −0.012 −0.022 −0.016 inf inf infGij −0.010 0.006 −0.017 −0.007 −0.014 418.601 infCar 0.042 −0.024 −0.017 −0.014 −0.015 0.001 infNe −0.012 −0.011 −0.030 −0.019 −0.026 −0.028 −0.015

17S.M. Francisco et al. / Journal of Experimental Marine Biology and Ecology 403 (2011) 14–20

3.2. Demographic analyses

The AMOVA results computed for both markers did not revealgenetic structure along the European coast (CR: FST=0.011,P=0.102; S7: FST=−0.007, P=0.708). For the CR the number ofmigrants (Nm) was always higher than 1, and the values of ΦST

(Table 2) and Corrected Average Differences (Table S2) were notsignificant for the great majority of comparisons among collectionssites. There were, however, some significant differences in bothmeasurements with some comparisons involving Cadiz (patternwithout geographical consistency). Considering the S7, no significantdifferences were found for the ΦST (Table 2) and Corrected AverageDifferences (Table S2).

The Mantel Test does not support a correlation between the Φst

and geographical distances for both the CR and S7 of the shanny, withthe coefficient of correlation itself being very close to zero (r=0.023P=0.342 and r=−0.182 P=0.737 respectively). Given the congru-ence of all the performed analyses, we decided to pool together the8 sampling sites as a European population for further analysis.

Neutrality tests yielded negative values for bothmarkers (Table 3);nevertheless, Fu's FS for the CR is the only significant value, suggestingdemographic expansion.

The analyses of the mismatch distributions for CR and S7 did notdiffer significantly from the model of sudden expansion for theEuropean population (Table 3, Fig. 3). For the CR (Fig. 3A) we candetect a stronger peak at 15 differences and another one around 3.With this marker the estimates point to an age of population between

400 Kya and 1 mya, and thus would not be the consequence of the lasttwo glaciations. The S7 graph (Fig. 3B) also revealed several peaks,with the highest being in the 7 differences. Even when we use this

Table 3Demographic parameters of Lipophrys pholis based CR and S7. Significant values ofprobability P are shown with an *. Neutrality tests: Fs (Fu's), D (Tajima's). Mismatchdistributions: τ (time in generations; upper and lower bounds of 95% CI in parenthesis),t (time in years), θ0 (theta0), N0 (female effective population size before the expansion),θ1 (theta1), N1 (female effective population size after the expansion), SSD (sum ofsquare deviation) and Hri (Harpending's Raggedness index). Estimates of populationparameters with LAMARC: θ (theta), Nf (female effective population size), g (growth rate),andN1% (age of population, accessed as the age atwhichNf drops below1%). Estimates oftMRCA (time to most recent common ancestor) with BEAST.

Population Europe

CR S7

Fs −24.040* −5.491D −1.136 −1.043

Mismatch distributionτ (95% CI) 14.764 7.607

(9.705–18.082) (0.291–57.273)t (95% CI) 388/1,079 ky 256 ky

(255/709–476/1,322) (10/1,927)θ0 0.004 0.002N0 105/292 34θ1 56.387 3.556N1 1,482,868/4,121,857 110,093SSD 0.007 0.081Hri 0.005 0.182

Lamarcθ (no growth) 0.0899 0.0007Nf (no growth) 899,374/2,498,261 15,967θ (with growth) 0.2227 –

Nf (with growth) 2,227,287/6,186,908 –

g 147.262 –

N1% 625/1,737 ky –

BeasttMRCA 1,022 ky 1,065 ky

Fig. 3. Mismatch distributions for the European population of Lipophrys pholis.A) Mitochondrial control region. B) First intron of the S7 ribosomal protein gene.

18 S.M. Francisco et al. / Journal of Experimental Marine Biology and Ecology 403 (2011) 14–20

marker with a much slower mutation rate, the age of the Europeanpopulation last expansion (84–98 kya) is older than the LGM. For bothmarkers N0 was unequivocally greater than zero.

The demographic parameters estimatedwith Lamarc for the CR arealso shown in Table 3. The corresponding analysis for S7 did not reachstability and was discarded. The confidence intervals of the age ofL. pholis population globally overlap with the previous analysis(625 kya–1,7 mya).

The results from the Bayesian skyline plot analyses are summarisedin Fig. 4. For the CR data, according to the plot (Fig. 4A) the shannypopulation experienced a faster population growth around 60 ky beforepresent. The skyline plot for S7 (Fig. 4B) identifies a moderate effectivepopulation size reduction between ~120–100 kya reaching a minimumof 600 individuals. A strong (~7-fold) and rapid size expansion began at~80 kya. It is noteworthy that time of the population reduction does notcorrelate with the LGM, and the expansion dates are not in agreementwith the end of the LGM, corroborating the previous analyses. Bothmarkers yielded congruent estimates for the tMRCA: ~1 My (Table 3).

4. Discussion

The present results raise at least two different issues: 1) the lack ofa detectable decline of genetic diversity in the north of the speciesrange and the homogeneity of the population along a large stretch ofEuropean coast; and 2) the complexity and depth of the genealogiesfound. We will address them in turn.

4.1. Homogeneity of the European population

The failure to detect strong fragmentation among Europeanlocations and the lack of support for a model of IBD lead us to assumethat the west European samples may be viewed as belonging to a

single population. It is interesting to note that there is no decline ofgenetic diversity with latitude.

In glacial peaks like the LGM, much of the North Sea was ice-covered and the waters north of Bay of Biscay were polar to sub-polar(Climap Project Members, 1984). Judging from the present distribu-tion of L. pholis, it is unlikely that it kept viable populations north ofSpain. In the meantime, habitats may have become available to thesouth, along northwest Africa and possibly in the Mediterranean.Indeed, the absence of stable populations in the present dayMediterranean may represent local extinction after the last glacialperiod, as the fish do not seem to be able to breed at temperatureshigher than 17 °C (Almada et al., 1990). The waters around Iberiawould be adequate for the survival of the species and it seemsreasonable to assume that Iberia must have played an important roleas a refugium from which the shanny recolonised West Europe whensea temperature rose in the last 10 thousand years (ky) [according toClimap Project Members (1984) the west Iberian coast yielded watertemperatures of 7.2 °C, figures comparable to the present day SST inthe North Sea]. If one considers a dispersal of 10 km in the samedirection per generation (which is a moderate estimate for marineorganisms with planktonic larvae) (Cowen and Sponaugle, 2009),10 ky would have ensured the recolonisation of the northern watersand would allow the northwards transport of much of the speciesgenetic diversity. If the scenario depicted above is correct, time todisperse from a south-western refugium is not the limiting factor for

Fig. 4. Extended Bayesian skyline plot for the European population of Lipophrys pholis.The solid line indicates the median estimate, whereas the dotted lines indicate the 95%credibility intervals. A) Mitochondrial control region. B) First intron of the S7 ribosomalprotein gene. Nef — female effective population size.

19S.M. Francisco et al. / Journal of Experimental Marine Biology and Ecology 403 (2011) 14–20

inshore benthic fish with planktonic larvae to recolonise west Europeafter a glaciation. Rather, the oceanographic conditions prevailingduring the breeding season and the behaviour and ecology from bothlarvae and adults will affect the colonisation process and the level ofpopulation differentiation.

4.2. The complexity and depth of the genealogies

All lines of available evidence point to the conclusion that the mainexpansion event recorded in the mismatch distribution occurred afteran earlier glaciation which apparently affected the population moreseverely than themost recent glacial maximum.While the smaller peak(three differences) in the mismatch distributions of the CR maycorrespond to an expansion after the LGM, the largest peak correspondsto 15 differences. Some lineages must even be older than that event,reaching 27 differences. The dates estimated with both markers areconsistently older than the LGM. The Bayesian skyline plots provideevidence for the main population expansions to have occurred duringthe Würm glacial cycle (12–110 kya; van Husen, 2004). Finally, thevalues of N0 are well above zero, indicating that the expansions beganfrom apopulation thatwas not subjected to a drastic bottleneck. Severalstudies on European coastal organisms also showed past expansionevents clearly older than the LGM: A. presbyter (Francisco et al., 2009),Pomatoschistusminutus (Larmuseau et al., 2009), Fucus serratus (Hoarauet al., 2007), Pollicipes pollicipes (Campo et al., 2010), etc. This holds even

for organisms that seem to have been able to invade the North Sea afterthe last glaciation (e.g. A. presbyter in Francisco et al., 2009; S. sprattus inDebes et al., 2008). This is not a generalised result and past expansionassociatedwith the LGMhas also been found in somemarine organisms(e.g. Coryphoblennius galerita in Domingues et al., 2007a,b;Necora puberin Sotelo et al., 2009; Palinurus elephas in Palero et al., 2008).

The haplotype network of the CR (Fig. 2a) includes a number ofstars of haplotypes at the tips of the branches of an older and complexdiversification process. Taken together, all analyses point to apopulation which retained substantial genetic diversity betweenglacial cycles and may coalesce after a genealogical history extendingto the middle or early Pleistocene. Several studies of temperatemarine fishes and other organisms show that the North-easternAtlantic frequently lacks strong phylogeographical structure, whilethe genealogies extend well behind the last glaciation (e.g. Debeset al., 2008).

We suggest that the populations of many marine organisms maymove up and down the west European shore with the glacial cyclesretaining much of their genetic diversity, although undergoingreductions and expansions in each glacial cycle. This will be especiallytrue for species with large population sizes and dispersal rates that aresufficiently high to enable them to track climate changes efficiently,by shifting their ranges. In such cases, a reduction of a very largepopulationmay still leave a large enough sample to retainmuch of theprevious genetic diversity during population size decreases.

Supplementarymaterials related to this article can be found onlineat doi: 10.1016/j.jembe.2011.03.020.

Acknowledgments

We acknowledge a grant to S.M.F. (SFRH/BD/12653/2003) fromFCT. The UIE-ISPA is funded by FCT through the pluri-annual andprogrammatic funding scheme (FEDER) as research unit #331/94. Thiswork was also funded by MARIN-ERA/MAR/0001/2008. The authorsare grateful to S. Chenu for her help with the laboratory work. [RH]

References

Almada, V.C., Barata, E.N., Gonçalves, E.J., Oliveira, R.F., 1990. On the breeding season ofLipophrys pholis (Pisces: Blenniidae) at Arrábida. Portugal. J. Mar. Biol. Assoc. U. K.70, 913–916.

Almada, V.C., Oliveira, R.F., Gonçalves, E.J., Almeida, A.J., Santos, R.S., Wirtz, P., 2001.Patterns of diversity of the north-eastern Atlantic blenniid fish fauna (Pisces:Blenniidae). Glob. Ecol. Biogeograph. 10, 411–422.

Almada, V.C., Pereira, A.M., Robalo, J.I., Fonseca, J.P., Levy, A., Maia, C., Valente, A., 2008.Mitochondrial DNA fails to reveal genetic structure in sea-lampreys along Europeanshores. Mol. Phylogenet. Evol. 46, 391–396.

Bernardi, G., Lape, J., 2005. Tempo andmode of speciation in the Baja California disjunctfish species Anisotremus davidsonii. Mol. Ecol. 14, 4085–4096.

Bowen, B.W., Muss, A., Rocha, L., Grant, W.S., 2006. Shallow mtDNA coalescence inAtlantic pygmy angelfishes (genus Centropyge) indicates a recent invasion from theIndian Ocean. J. Hered. 97, 1–12.

Bremer, J.R.A., Viñas, J., Mejuto, J., Ely, B., Pla, C., 2005. Comparative phylogeography ofAtlantic blue fin tuna and swordfish: the combined effects of vicariance, secondarycontact, introgression, and population expansion on the regional phylogenies oftwo highly migratory pelagic fishes. Mol. Phylogenet. Evol. 36, 169–187.

Briggs, J.V., 1996. Global Biogeography. Developments in paleontology and stratigra-phy, vol.14. Elsevier, Amsterdam.

Campo, D., Molares, J., Garcia, L., Fernandez-Rueda, P., Garcia-Gonzalez, C., Garcia-Vazquez, E., 2010. Phylogeography of the European stalked barnacle (Pollicipespollicipes): identification of glacial refugia.

Chevolot, M., Wolfs, P.H.J., Pálsoon, J., Rijnsdorp, A.D., Stam, W.T., Olsen, J.L., 2006.Population structure and historical demography of the thorny skate (Amblyrajaradiate, Rajidae) in North Atlantic. Mar. Biol. 151, 1275–1286.

Chow, S., Hazama, K., 1998. Universal PCR primers for S7 ribosomal protein gene intronsin fish. Mol. Ecol. 7, 1247–1263.

Clement, M., Posada, D., Crandall, K.A., 2000. TCS: a computer program to estimate genegenealogies. Mol. Ecol. 9, 1657–1660.

Climap Project Members, 1984. The last interglacial ocean. Quat. Res. 21, 123–224.Cowen, R.K., Sponaugle, S., 2009. Larval dispersal and marine population connectivity.

Ann. Rev. Mar. Sci. 1, 443–466.Debes, P.V., Zachos, F.E., Hanel, R., 2008. Mitochondrial phylogeography of the European

sprat (Sprattus sprattus L., Clupeidae) reveals isolated climatically vulnerablepopulations in the Mediterranean Sea and range expansion in the northeastAtlantic. Mol. Ecol. 17, 3873–3888.

20 S.M. Francisco et al. / Journal of Experimental Marine Biology and Ecology 403 (2011) 14–20

Domingues, V.S., Bucciarelli, G., Almada, V.C., Bernardi, G., 2005. Historical colonizationand demography of the Mediterranean damselfish Chromis chromis. Mol. Ecol. 13,4051–4063.

Domingues, V.S., Faria, C., Stefanni, S., Satos, R.S., Brito, A., Almada, V.C., 2007a. Geneticdivergence in the Atalntic Mediterranean Montagu's blenny Coryphoblenniusgalerita (Linnaeus 1758) revealed by molecular and morphological characters.Mol. Ecol. 16, 3592–3605.

Domingues, V.S., Santos, R.S., Brito, A., Alexandrou, M., Almada, V.C., 2007b.Mitochondrialand nuclearmarkers reveal isolation-by-distance andeffects of Pleistoceneglaciationsin the north-eastern Atlantic and Mediterranean populations of the white seabream(Diplodus sargus, L.). J. Exp. Mar. Biol. Ecol. 346, 102–113.

Drummond, A.J., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis by samplingtrees. BMC Evol. Biol. 7, 214.

Excoffier, L., Smouse, P.E., Quattro, J.M., 1992. Analysis of molecular variance inferredfrom metric distances among DNA haplotypes: application to human mitochon-drial DNA restriction data. Genetics 131, 479–491.

Excoffier, L., Laval, G., Schneider, S., 2005. Arlequin ver 3.0: An integrated softwarepackage for population genetics data analysis. Evol. Bioinform. Online 1, 47–50.

Faria, C., Almada, V.C., Gonçalves, E.J., 1996. Juvenile recruitment, growth andmaturation of Lipophrys pholis (Pisces: Blenniidae), from the west coast of Portugal.J. Fish Biol. 49, 727–730.

Faria, C., Borges, R., Gil, F., Almada, V.C., Gonçalves, E.J., 2002. Embryonic and larvaldevelopment of Lipophrys pholis (Pisces: Blenniidae). Sci. Mar. 66, 21–26.

Francisco, S.M., Vieira, M.N., Almada, V.C., 2006. Genetic structure and historicaldemography of the shanny Lipophrys pholis in the Portuguese coast based onmitochondrial DNA analysis. Mol. Phylogenet. Evol. 39, 288–292.

Francisco, S.M., Castilho, R., Soares, M., Congiu, L., Brito, A., Vieira, M.N., Almada, V.C.,2009. Phylogeography and demographic history of Atherina presbyter (Pisces:Atherinidae) in the North-eastern Atlantic based on mitochondrial DNA. Mar. Biol.156, 1421–1432.

Fu, Y.X., 1997. Statistical tests of neutrality of mutations against population growth,hitchhiking and background selection. Genetics 147, 915–925.

Gysels, E.S., Hellemans, B., Pampoulie, C., Volckaert, A.M., 2004. Phylogeography of thecommon goby, Pomatoschistus microps, with particular emphasis on the coloniza-tion of the Mediterranean and the North Sea. Mol. Ecol. 13, 403–417.

Heled, J., Drummond, A.J., 2006. Bayesian inferecence of population size history frommultiple loci. BMC Evol. Biol. 8, 289.

Hoarau, G., Coyer, J.A., Veldsink, J.H., Stam, W.T., Olsen, J.L., 2007. Glacial refugia andrecolonization pathways in the brown seaweed Fucus serratus. Mol. Ecol. 16, 3606–3616.

Kuhner, M.K., 2006. LAMARC 2.0: Maximum likelihood and Bayesian estimations ofpopulation parameters. Bioinformatics 22, 768–770.

Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, H.,Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., Thompson, J.D., Gibson, T.J., Higgins,D.G., 2007. Clustal W and Clustal X version 2.0. Bioinformatics 23, 2947–2948.

Larmuseau, M.H.D., Van Houdt, J.K.J., Guelinckx, J., Hellemans, B., Volckaert, F.A.M.,2009. Distributional and demographic consequences of Pleistocene climatefluctuations for a marine demersal fish in the north-eastern Atlantic. J. Biogeo.36, 1138–1151.

Mäkinen, H.S., Merilä, J., 2008. Mitochondrial DNA phylogeography of the three-spinedstickleback (Gasterosteus aculeatus) in Europe — Evidence for multiple glacialrefugia. Mol. Phylogenet. Evol. 46, 167–182.

Mantel, N., 1967. The detection of disease clustering and a generalized regressionapproach. Cancer Res. 27, 209–220.

Milton, P., 1983. Biology of littoral blenniid fishes on the coast of South-West England.Symp. Zool. Soc. Lond. 44, 263–306.

Ostellari, L., Barglloni, L., Penzo, E., Patarnello, P., Patarnello, T., 1996. Optimization ofsingle-strand conformation polymorphism and sequence analysis of the mito-chondrial control region in Pagellus bogaraveo (Sparidae, Teleostei): rationalizedtools in fish population biology. Anim. Genet. 27, 423–427.

Palero, F., Abelló, P., Macpherson, E., Gristina, M., Pascual, M., 2008. Phylogeography ofthe European spiny lobster (Palinurus elephas): influence of current oceanograph-ical features and historical processes. Mol. Phylogen. Evol. 48, 708–717.

Rambaut, A., Drummond, A.J., 2007. Tracer [computer program]. http://tree.bio.ed.ac.uk/software/tracer/.

Rogers, A.R., 1995. Genetic evidence for a Pleistocene population explosion. Evolution49, 608–615.

Rogers, A.R., Harpensding, H., 1992. Population growth makes waves in the distributionof pairwise genetic differences. Mol. Biol. Evol. 9, 552–569.

Smouse, P.E., Long, J.C., Sokal, R.R., 1986. Multiple regression and correlation extensionsof the Mantel Test of matrix correspondence. Syst. Zool. 35, 627–632.

Sotelo, G., Posada, D., Morán, P., 2009. Low-mitochondrial diversity and lack of structurein the velvet swimming crab Necora puber along the Galician coast. Mar Biol 156,1039–1048.

Sousa-Santos, C., Robalo, J.I., Collares-Pereira, M.J., Almada, V.C., 2005. Heterozygousindels as useful tools in the reconstruction of DNA sequences and in assessment ofploidy level and genomic constitution of hybrid organisms. DNA Seq. 16, 462–467.

Stefanni, S., Domingues, V., Bouton, N., Santos, R.S., Almada, F., Almada, V., 2006.Phylogeny of the shanny, Lipophrys pholis, from the NE Atlantic usingmitochondrialDNA markers. Mol. Phylogenet. Evol. 39, 282–287.

Swofford, D.L., 2000. PAUP* Phylogenetic Analysis Using Parsimony (*and othermethods), Version 4. Sinauer Associates Inc, Sunderland, MA, USA.

Tajima, F., 1989. The effect of change in population size on DNA polymorphism.Genetics 123, 597–660.

Templeton, A.R., Crandall, K.A., Sing, C.F., 1992. A cladistic-analysis of phenotypicassociations with haplotypes inferred from restriction endonuclease mapping andDNA-sequence data. III. Cladogram estimation. Genetics 132, 619–633.

Van Husen, D., 2004. Quaternary glaciations in Austria. In: Ehlers, J., Gibbard, P.L. (Eds.),Quaternary glaciations — Extent and Chronology. Part I. Elsevier, Europe, pp. 1–14.